{"id":7459,"date":"2026-08-10T09:44:23","date_gmt":"2026-08-10T09:44:23","guid":{"rendered":"https:\/\/nokobox.com\/index.php\/item\/oreilly-deep-learning-patterns-and-practices-2022-3\/"},"modified":"2026-08-10T09:44:23","modified_gmt":"2026-08-10T09:44:23","slug":"oreilly-deep-learning-patterns-and-practices-2022-3","status":"publish","type":"digital_item","link":"https:\/\/nokobox.com\/index.php\/item\/oreilly-deep-learning-patterns-and-practices-2022-3\/","title":{"rendered":"Oreilly \u2013 Deep Learning Patterns and Practices 2022-3"},"content":{"rendered":"<div class=\"w-post-elm post_content\">\n<h2><span style=\"vertical-align: inherit\">Description<\/span><\/h2>\n<p data-sourcepos=\"7:1-7:179\"><span style=\"vertical-align: inherit\">Deep Learning Patterns and Practices is a comprehensive guide that introduces you to best practices, repeatable architectures, and design patterns to take your deep learning models from testing to production. A major challenge in deep learning is moving emerging technologies from R&amp;D labs to production. This book helps you overcome this challenge with the latest insights from author Andrew Frelitsch, a fellow at Google Cloud AI. Deep learning models are presented in a unique new way as extensible design patterns that you can easily use in your software projects. Each valuable technique is presented in simple language, accompanied by easy-to-understand diagrams and code samples.<\/span><\/p>\n<h3 data-sourcepos=\"9:1-9:25\"><span style=\"vertical-align: inherit\">What you will learn:<\/span><\/h3>\n<ul data-sourcepos=\"11:1-17:0\">\n<li data-sourcepos=\"11:1-11:120\"><span style=\"vertical-align: inherit\">The topic of modern convolutional networks: You will become familiar with the internal details of these networks and understand their performance well.<\/span><\/li>\n<li data-sourcepos=\"12:1-12:170\"><span style=\"vertical-align: inherit\">Procedural Reuse Design Pattern for CNN Architectures: With this pattern, you can design your convolutional network architectures in a modular and reusable way.<\/span><\/li>\n<li data-sourcepos=\"13:1-13:138\"><span style=\"vertical-align: inherit\">Models suitable for mobile and IoT devices: You will learn how to optimize deep learning models for devices with limited resources.<\/span><\/li>\n<li data-sourcepos=\"14:1-14:118\"><span style=\"vertical-align: inherit\">Deploying Large-Scale Models: You will learn about different methods for deploying and managing large-scale deep learning models.<\/span><\/li>\n<li data-sourcepos=\"15:1-15:133\"><span style=\"vertical-align: inherit\">Optimizing Hyperparameter Settings: You will learn how to optimize the hyperparameter settings of your model to achieve the best performance.<\/span><\/li>\n<li data-sourcepos=\"16:1-17:0\"><span style=\"vertical-align: inherit\">Migrating a model to a production environment: You will learn about the different steps involved in migrating a model from a development environment to a production environment.<\/span><\/li>\n<\/ul>\n<h3 data-sourcepos=\"18:1-18:33\"><span style=\"vertical-align: inherit\">This course is suitable for people who:<\/span><\/h3>\n<ul data-sourcepos=\"20:1-23:0\">\n<li data-sourcepos=\"20:1-20:43\"><span style=\"vertical-align: inherit\">They are familiar with the Python programming language.<\/span><\/li>\n<li data-sourcepos=\"21:1-21:42\"><span style=\"vertical-align: inherit\">Are familiar with the basic concepts of deep learning.<\/span><\/li>\n<li data-sourcepos=\"22:1-23:0\"><span style=\"vertical-align: inherit\">They are looking to increase their skills in deep learning and model deployment.<\/span><\/li>\n<\/ul>\n<h3><span style=\"vertical-align: inherit\">Course details: Deep Learning Patterns and Practices<\/span><\/h3>\n<ul>\n<li><span style=\"vertical-align: inherit\">Publisher: <\/span><a href=\"https:\/\/href.li\/?https:\/\/www.oreilly.com\/library\/view\/deep-learning-patterns\/9781617298264\/\/\" target=\"_blank\" rel=\"noopener\"><span style=\"vertical-align: inherit\">Oreilly<\/span><\/a><\/li>\n<li><span style=\"vertical-align: inherit\">Instructor: <\/span><a href=\"https:\/\/downloadlynet.ir\/tag\/andrew-ferlitsch\/\"><span style=\"vertical-align: inherit\">Andrew Ferlitsch<\/span><\/a><\/li>\n<li><span style=\"vertical-align: inherit\">Training level: Beginner to advanced<\/span><\/li>\n<li><span style=\"vertical-align: inherit\">Training duration: 13 hours and 53 minutes<\/span><\/li>\n<\/ul>\n<h3><span style=\"vertical-align: inherit\">Course headings<\/span><\/h3>\n<ul>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">Part 1. Deep learning fundamentals<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">1 Designing modern machine learning<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">1.1 A focus on adaptability<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">1.1.1 Computer vision leading the way<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">1.1.2 Beyond computer vision: NLP, NLU, structured data<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">1.2 The evolution in machine learning approaches<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">1.2.1 Classical AI vs. narrow AI<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">1.2.2 Next steps in computer learning<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">1.3 The benefits of design patterns<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">Summary<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">2 Deep neural networks<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">2.1 Neural network basics<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">2.1.1 Input layer<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">2.1.2 Deep neural networks<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">2.1.3 Feed-forward networks<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">2.1.4 Sequential API method<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">2.1.5 Functional API methods<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">2.1.6 Input shape vs. input layer<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">2.1.7 Dense layer<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">2.1.8 Activation functions<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">2.1.9 Shorthand syntax<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">2.1.10 Improving accuracy with an optimizer<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">2.2 DNN binary classifier<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">2.3 DNN multiclass classifier<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">2.4 DNN multilabel multiclass classifier<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">2.5 Simple image classifier<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">2.5.1 Flattening<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">2.5.2 Overfitting and dropout<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">Summary<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">3 Convolutional and residual neural networks<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">3.1 Convolutional neural networks<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">3.1.1 Why we use a CNN over a DNN for image models<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">3.1.2 Downsampling (resizing)<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">3.1.3 Feature detection<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">3.1.4 Pooling<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">3.1.5 Flattening<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">3.2 The ConvNet design for a CNN<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">3.3 VGG networks<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">3.4 ResNet networks<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">3.4.1 Architecture<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">3.4.2 Batch normalization<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">3.4.3 ResNet50<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">Summary<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">4 Training fundamentals<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">4.1 Forward feeding and backward propagation<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">4.1.1 Feeding<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">4.1.2 Backward propagation<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">4.2 Dataset splitting<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">4.2.1 Training and test sets<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">4.2.2 One-hot encoding<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">4.3 Data normalization<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">4.3.1 Normalization<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">4.3.2 Standardization<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">4.4 Validation and overfitting<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">4.4.1 Validation<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">4.4.2 Loss monitoring<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">4.4.3 Going deeper with layers<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">4.5 Convergence<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">4.6 Checkpointing and early stopping<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">4.6.1 Checkpointing<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">4.6.2 Early stopping<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">4.7 Hyperparameters<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">4.7.1 Epochs<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">4.7.2 Steps<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">4.7.3 Batch size<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">4.7.4 Learning rate<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">4.8 Invariance<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">4.8.1 Translational invariance<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">4.8.2 Scale invariance<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">4.8.3 TF.Keras ImageDataGenerator<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">4.9 Raw (disk) datasets<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">4.9.1 Directory structure<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">4.9.2 CSV file<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">4.9.3 JSON file<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">4.9.4 Reading images<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">4.9.5 Resizing<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">4.10 Model save\/restore<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">4.10.1 Save<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">4.10.2 Restore<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">Summary<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">Part 2. Basic design pattern<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">5 Procedural design patterns<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">5.1 Basic neural network architecture<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">5.2 Stem component<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">5.2.1 VGG<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">5.2.2 ResNet<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">5.2.3 ResNeXt<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">5.2.4 Xception<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">5.3 Pre-stem<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">5.4 Learner component<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">5.4.1 ResNet<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">5.4.2 DenseNet<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">5.5 Task component<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">5.5.1 ResNet<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">5.5.2 Multilayer output<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">5.5.3 SqueezeNet<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">5.6 Beyond computer vision: NLP<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">5.6.1 Natural-language understanding<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">5.6.2 Transformer architecture<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">Summary<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">6 Wide convolutional neural networks<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">6.1 Inception v1<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">6.1.1 Naive inception module<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">6.1.2 Inception v1 module<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">6.1.3 Stem<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">6.1.4 Learner<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">6.1.5 Auxiliary classifiers<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">6.1.6 Classifier<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">6.2 Inception v2: Factoring convolutions<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">6.3 Inception v3: Architecture redesign<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">6.3.1 Inception groups and blocks<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">6.3.2 Normal convolution<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">6.3.3 Spatial separable convolution<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">6.3.4 Stem redesign and implementation<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">6.3.5 Auxiliary classifier<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">6.4 ResNeXt: Wide residual neural networks<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">6.4.1 ResNeXt block<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">6.4.2 ResNeXt architecture<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">6.5 Wide residual network<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">6.5.1 WRN-50-2 architecture<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">6.5.2 Wide residual block<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">6.6 Beyond computer vision: Structured data<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">Summary<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">7 Alternative connectivity patterns<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">7.1 DenseNet: Densely connected convolutional neural network<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">7.1.1 Dense group<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">7.1.2 Dense block<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">7.1.3 DenseNet macro-architecture<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">7.1.4 Dense transition block<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">7.2 Xception: Extreme Inception<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">7.2.1 Xception architecture<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">7.2.2 Entry flow of Xception<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">7.2.3 Middle flow of Xception<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">7.2.4 Exit flow of Xception<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">7.2.5 Depthwise separable convolution<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\">7.2.6 Depthwise convolution<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">7.2.7 Pointwise convolution<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">7.3 SE-Net: Squeeze and excitation<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">7.3.1 Architecture of SE-Net<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">7.3.2 Group and block of SE-Net<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">7.3.3 SE link<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">Summary<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">8 Mobile convolutional neural networks<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">8.1 MobileNet v1<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">8.1.1 Architecture<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">8.1.2 Width multiplier<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">8.1.3 Resolution multiplier<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">8.1.4 Stem<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">8.1.5 Learner<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">8.1.6 Classifier<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">8.2 MobileNet v2<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">8.2.1 Architecture<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">8.2.2 Stem<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">8.2.3 Learner<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">8.2.4 Classifier<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">8.3 SqueezeNet<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">8.3.1 Architecture<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">8.3.2 Stem<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">8.3.3 Learner<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">8.3.4 Classifier<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">8.3.5 Bypass connections<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">8.4 ShuffleNet v1<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">8.4.1 Architecture<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">8.4.2 Stem<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">8.4.3 Learner<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">8.5 Deployment<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">8.5.1 Quantization<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">8.5.2 TF Lite conversion and prediction<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">Summary<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">9 Autoencoders<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">9.1 Deep neural network autoencoders<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">9.1.1 Autoencoder architecture<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">9.1.2 Encoder<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">9.1.3 Decoder<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">9.1.4 Training<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">9.2 Convolutional autoencoders<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">9.2.1 Architecture<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">9.2.2 Encoder<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">9.2.3 Decoder<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">9.3 Sparse autoencoders<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">9.4 Denoising autoencoders<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">9.5 Super-resolution<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">9.5.1 Pre-upsampling SR<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">9.5.2 Post-upsampling SR<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">9.6 Pretext tasks<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">9.7 Beyond computer vision: sequence to sequence<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">Summary<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">Part 3. Working with pipelines<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">10 Hyperparameter tuning<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">10.1 Weight initialization<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">10.1.1 Weight distributions<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">10.1.2 Lottery hypothesis<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">10.1.3 Warm-up (numerical stability)<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">10.2 Hyperparameter search fundamentals<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">10.2.1 Manual method for hyperparameter search<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">10.2.2 Grid search<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">10.2.3 Random search<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">10.2.4 KerasTuner<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">10.3 Learning rate scheduler<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">10.3.1 Keras decay parameter<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">10.3.2 Keras learning rate scheduler<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">10.3.3 Ramp<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">10.3.4 Constant step<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">10.3.5 Cosine annealing<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">10.4 Regularization<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">10.4.1 Weight regularization<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">10.4.2 Label smoothing<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">10.5 Beyond computer vision<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">Summary<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">11 Transfer learning<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">11.1 TF.Keras prebuilt models<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">11.1.1 Base model<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">11.1.2 Pretrained ImageNet models for prediction<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">11.1.3 New classifier<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">11.2 TF Hub prebuilt models<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">11.2.1 Using TF Hub pretrained models<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">11.2.2 New classifier<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">11.3 Transfer learning between domains<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">11.3.1 Similar tasks<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">11.3.2 Distinct tasks<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">11.3.3 Domain-specific weights<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">11.3.4 Domain transfer weight initialization<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">11.3.5 Negative transfer<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">11.4 Beyond computer vision<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">Summary<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">12 Data distributions<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">12.1 Distribution types<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">12.1.1 Population distribution<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">12.1.2 Sampling distribution<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">12.1.3 Subpopulation distribution<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">12.2 Out of distribution<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">12.2.1 The MNIST curated dataset<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">12.2.2 Setting up the environment<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">12.2.3 The challenge (\u201cin the wild\u201d)<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">12.2.4 Training as a DNN<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">12.2.5 Training as a CNN<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">12.2.6 Image augmentation<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">12.2.7 Final test<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">Summary<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">13 Data pipeline<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">13.1 Data formats and storage<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">13.1.1 Compressed and raw-image formats<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">13.1.2 HDF5 format<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">13.1.3 DICOM format<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">13.1.4 TFRecord format<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">13.2 Data feeding<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">13.2.1 NumPy<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">13.2.2 TFRecord<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">13.3 Data preprocessing<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">13.3.1 Preprocessing with a pre-stem<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">13.3.2 Preprocessing with TF Extended<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">13.4 Data augmentation<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">13.4.1 Invariance<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">13.4.2 Augmentation with tf.data<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">13.4.3 Pre-stem<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">Summary<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">14 Training and deployment pipeline<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">14.1 Model feeding<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">14.1.1 Model feeding with tf.data.Dataset<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">14.1.2 Distributed feeding with tf.Strategy<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">14.1.3 Model feeding with TFX<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">14.2 Training schedulers<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">14.2.1 Pipeline versioning<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">14.2.2 Metadata<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">14.2.3 History<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">14.3 Model evaluations<\/li>\n<li dir=\"ltr\" style=\"text-align: left\">14.3.1 Candidate vs. blessed model<\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">14.3.2 TFX evaluation<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">14.4 Serving predictions<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">14.4.1 On-demand (live) serving<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">14.4.2 Batch prediction<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">14.4.3 TFX pipeline components for deployment<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">14.4.4 A\/B testing<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">14.4.5 Load balancing<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">14.4.6 Continuous evaluation<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">14.5 Evolution in production pipeline design<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">14.5.1 Machine learning as a pipeline<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">14.5.2 Machine learning as a CI\/CD production process<\/span><\/li>\n<li dir=\"ltr\" style=\"text-align: left\"><span style=\"vertical-align: inherit\">14.5.3 Model amalgamation in production<\/span><\/li>\n<\/ul>\n<h3><span style=\"vertical-align: inherit\">Images from the Deep Learning Patterns and Practices course<\/span><\/h3>\n<h2><img loading=\"lazy\" decoding=\"async\" class=\"aligncenter wp-image-949525 size-full\" src=\"https:\/\/downloadly.ir\/wp-content\/uploads\/2024\/12\/Deep-Learning-Patterns-and-Practices1.png\" alt=\"Deep Learning Patterns and Practices\" width=\"1031\" height=\"330\"><\/h2>\n<h3><span style=\"vertical-align: inherit\">Sample course video<\/span><\/h3>\n<div style=\"width: 640px;\" class=\"wp-video\"><span class=\"mejs-offscreen\">Video Player<\/span><\/p>\n<div id=\"mep_0\" class=\"mejs-container mejs-container-keyboard-inactive wp-video-shortcode mejs-video\" tabindex=\"0\" role=\"application\" aria-label=\"Video Player\" style=\"width: 640px; height: 360px; min-width: 217px;\">\n<div class=\"mejs-inner\">\n<div class=\"mejs-mediaelement\"><mediaelementwrapper id=\"video-148355-1\"><video class=\"wp-video-shortcode\" id=\"video-148355-1_html5\" width=\"640\" height=\"360\" preload=\"metadata\" src=\"https:\/\/dl.downloadly.ir\/Files\/Elearning\/Sample\/Deep_Learning_Patterns_and_Practices_Downloadly.ir.mp4?_=1\" style=\"width: 640px; height: 360px;\"><source type=\"video\/mp4\" src=\"https:\/\/dl.downloadly.ir\/Files\/Elearning\/Sample\/Deep_Learning_Patterns_and_Practices_Downloadly.ir.mp4?_=1\"><a href=\"https:\/\/dl.downloadly.ir\/Files\/Elearning\/Sample\/Deep_Learning_Patterns_and_Practices_Downloadly.ir.mp4?nocache=1786109774947\">https:\/\/dl.downloadly.ir\/Files\/Elearning\/Sample\/Deep_Learning_Patterns_and_Practices_Downloadly.ir.mp4<\/a><\/video><\/mediaelementwrapper><\/div>\n<div class=\"mejs-layers\">\n<div class=\"mejs-poster mejs-layer\" style=\"display: none; width: 100%; height: 100%;\"><\/div>\n<div class=\"mejs-overlay mejs-layer\" style=\"display: none; width: 100%; height: 100%;\">\n<div class=\"mejs-overlay-loading\"><span class=\"mejs-overlay-loading-bg-img\"><\/span><\/div>\n<\/div>\n<div class=\"mejs-overlay mejs-layer\" style=\"display: none; width: 100%; height: 100%;\">\n<div class=\"mejs-overlay-error\"><\/div>\n<\/div>\n<div class=\"mejs-overlay mejs-layer mejs-overlay-play\" style=\"width: 100%; height: 100%;\">\n<div class=\"mejs-overlay-button\" role=\"button\" tabindex=\"0\" aria-label=\"Play\" aria-pressed=\"false\"><\/div>\n<\/div>\n<\/div>\n<div class=\"mejs-controls\">\n<div class=\"mejs-button mejs-playpause-button mejs-play\"><button type=\"button\" aria-controls=\"mep_0\" title=\"Play\" aria-label=\"Play\" tabindex=\"0\"><\/button><\/div>\n<div class=\"mejs-time mejs-currenttime-container\" role=\"timer\" aria-live=\"off\"><span class=\"mejs-currenttime\">00:00<\/span><\/div>\n<div class=\"mejs-time-rail\"><span class=\"mejs-time-total mejs-time-slider\" role=\"slider\" tabindex=\"0\" aria-label=\"Time Slider\" aria-valuemin=\"0\" aria-valuemax=\"0\" aria-valuenow=\"0\" aria-valuetext=\"00:00\"><span class=\"mejs-time-buffering\" style=\"display: none;\"><\/span><span class=\"mejs-time-loaded\"><\/span><span class=\"mejs-time-current\"><\/span><span class=\"mejs-time-hovered no-hover\"><\/span><span class=\"mejs-time-handle\"><span class=\"mejs-time-handle-content\"><\/span><\/span><span class=\"mejs-time-float\"><span class=\"mejs-time-float-current\">00:00<\/span><span class=\"mejs-time-float-corner\"><\/span><\/span><\/span><\/div>\n<div class=\"mejs-time mejs-duration-container\"><span class=\"mejs-duration\">00:00<\/span><\/div>\n<div class=\"mejs-button mejs-volume-button mejs-mute\"><button type=\"button\" aria-controls=\"mep_0\" title=\"Mute\" aria-label=\"Mute\" tabindex=\"0\"><\/button><a href=\"javascript:void(0);\" class=\"mejs-volume-slider\" aria-label=\"Volume Slider\" aria-valuemin=\"0\" aria-valuemax=\"100\" role=\"slider\" aria-orientation=\"vertical\"><span class=\"mejs-offscreen\">Use Up\/Down Arrow keys to increase or decrease volume.<\/span><\/p>\n<div class=\"mejs-volume-total\">\n<div class=\"mejs-volume-current\" style=\"bottom: 0px; height: 100%;\"><\/div>\n<div class=\"mejs-volume-handle\" style=\"bottom: 100%; margin-bottom: -3px;\"><\/div>\n<\/div>\n<p><\/a><\/div>\n<div class=\"mejs-button mejs-fullscreen-button\"><button type=\"button\" aria-controls=\"mep_0\" title=\"Fullscreen\" aria-label=\"Fullscreen\" tabindex=\"0\"><\/button><\/div>\n<\/div>\n<\/div>\n<\/div>\n<\/div>\n<div>\n<h3><span style=\"vertical-align: inherit\">Installation Guide<\/span><\/h3>\n<p><span style=\"vertical-align: inherit\">After Extract, view with your favorite player.<\/span><\/p>\n<p><span style=\"vertical-align: inherit\">Subtitles: None<\/span><\/p>\n<p><span style=\"vertical-align: inherit\">Quality: 720p<\/span><\/p>\n<\/div>\n<h3><span style=\"vertical-align: inherit\">Download link<\/span><\/h3>\n<p><a href=\"https:\/\/dl2.downloadly.ir\/Files\/Elearning\/Oreilly_Deep_Learning_Patterns_and_Practices_2022-3.part1_Downloadly.ir.rar?nocache=1786109772\"><span style=\"vertical-align: inherit\">Download Part 1 \u2013 1 GB<\/span><\/a><\/p>\n<p><a href=\"https:\/\/dl2.downloadly.ir\/Files\/Elearning\/Oreilly_Deep_Learning_Patterns_and_Practices_2022-3.part2_Downloadly.ir.rar?nocache=1786109772\"><span style=\"vertical-align: inherit\">Download Part 2 \u2013 552 MB<\/span><\/a><\/p>\n<p><span style=\"vertical-align: inherit\">File(s) password: www.downloadly.ir<\/span><\/p>\n<h3><span style=\"vertical-align: inherit\">File size<\/span><\/h3>\n<p><span style=\"vertical-align: inherit\">1.5 GB<\/span><\/p>\n<\/div>\n","protected":false},"excerpt":{"rendered":"<p>Description Deep Learning Patterns and Practices is a comprehensive guide that introduces you to best practices, repeatable architectures, and design patterns t<\/p>\n","protected":false},"author":1,"template":"","dgi_category":[10458],"dgi_tag":[68129,68130,68131,68132,68133,68134],"class_list":["post-7459","digital_item","type-digital_item","status-publish","has-post-thumbnail","hentry","dgi_category-video-tutorials","dgi_tag-andrew-ferlitsch","dgi_tag-course-deep-learning-patterns-and-practices","dgi_tag-download-course-deep-learning-patterns-and-practices","dgi_tag-download-deep-learning-patterns-and-practices","dgi_tag-free-deep-learning-patterns-and-practices","dgi_tag-free-download-deep-learning-patterns-and-practices"],"_links":{"self":[{"href":"https:\/\/nokobox.com\/index.php\/wp-json\/wp\/v2\/digital_item\/7459","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/nokobox.com\/index.php\/wp-json\/wp\/v2\/digital_item"}],"about":[{"href":"https:\/\/nokobox.com\/index.php\/wp-json\/wp\/v2\/types\/digital_item"}],"author":[{"embeddable":true,"href":"https:\/\/nokobox.com\/index.php\/wp-json\/wp\/v2\/users\/1"}],"version-history":[{"count":0,"href":"https:\/\/nokobox.com\/index.php\/wp-json\/wp\/v2\/digital_item\/7459\/revisions"}],"wp:attachment":[{"href":"https:\/\/nokobox.com\/index.php\/wp-json\/wp\/v2\/media?parent=7459"}],"wp:term":[{"taxonomy":"dgi_category","embeddable":true,"href":"https:\/\/nokobox.com\/index.php\/wp-json\/wp\/v2\/dgi_category?post=7459"},{"taxonomy":"dgi_tag","embeddable":true,"href":"https:\/\/nokobox.com\/index.php\/wp-json\/wp\/v2\/dgi_tag?post=7459"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}